Search papers, labs, and topics across Lattice.
The paper introduces SLER-IR, a novel framework for all-in-one image restoration that uses spherical layer-wise expert routing to dynamically activate specialized experts across network layers. To improve routing reliability, they employ Spherical Uniform Degradation Embedding with contrastive learning, mapping degradation representations onto a hypersphere. The framework also incorporates a Global-Local Granularity Fusion (GLGF) module to handle spatially non-uniform degradations, achieving state-of-the-art performance on multi-task image restoration benchmarks.
Achieve state-of-the-art image restoration by routing inputs through a dynamically-selected mixture of experts, guided by a spherical embedding of degradation type.
Image restoration under diverse degradations remains challenging for unified all-in-one frameworks due to feature interference and insufficient expert specialization. We propose SLER-IR, a spherical layer-wise expert routing framework that dynamically activates specialized experts across network layers. To ensure reliable routing, we introduce a Spherical Uniform Degradation Embedding with contrastive learning, which maps degradation representations onto a hypersphere to eliminate geometry bias in linear embedding spaces. In addition, a Global-Local Granularity Fusion (GLGF) module integrates global semantics and local degradation cues to address spatially non-uniform degradations and the train-test granularity gap. Experiments on three-task and five-task benchmarks demonstrate that SLER-IR achieves consistent improvements over state-of-the-art methods in both PSNR and SSIM. Code and models will be publicly released.
